In order to count the number of configured inference processors, the inference processor factory instantiates all ingest pipelines. This causes problems with processors whose factories have not yet been fully configured. It would also be preferable if there were a lighter-weight method for obtaining the count of inference processors.